Supporting Asynchronous Collaboration for Interactive Visualization.

Interactive visualizations leverage human visual processing to increase the scale of information with which we can effectively work. However, most visualization research to date relies on a single-user model, overlooking the social nature of visual media. Visualizations are used not only to explore and analyze data, but to communicate findings. People may disagree on how to interpret data and contribute contextual knowledge. Furthermore, some data sets are so large that thorough exploration by a single person is unlikely. Such scenarios arise regularly in scientific collaboration, business intelligence, and public data consumption. This thesis recasts interactive visualizations as not just analytic tools, but social spaces supporting collective data analysis. To this aim, I introduce theoretical design considerations guiding the invention of social visual analysis tools and present the design, implementation, and evaluation of interactive systems based on these principles. The first such system is sense.us, a web site supporting asynchronous collaboration across a variety of visualization types. The site supports view sharing, discussion, graphical annotation, and social navigation and includes novel interaction elements. User studies of the system reveal emergent patterns of social data analysis, cycles of observation and hypothesis, and the complementary roles of social navigation and data-driven exploration. Based on design considerations and lessons learned from sense.us, this dissertation also introduces new techniques to support collaborative interaction around visualizations. The scented widgets system embeds visualizations of social activity in common user interface controls to enhance collective information foraging. A generalized selection framework represents collaborative annotations as declarative queries over visualized data, enabling annotation of dynamic data across multiple visualization views. Interactive query relaxation enables users to further generalize selections along data dimensions of interest. New graphical histories for visualization support analysis and accelerate collaborative sharing of findings, and a framework for animated transitions better communicates the relationship between views in an analysis session. As evidenced in a series of evaluative studies, these components enable teams to collaborate more effectively as they conduct visual data analysis.